PV monitoring optimizer¶

This document contains the description for the PV portfolio optimizer. In the first stage, this optimizer will clasify the PV systems within a PV portfolio into 3 main categories:

Green: if the acumulated power losses of a PV system are not more than 50% of 1 day man-labor cost.
Orange: if the acumulated power losses of a PV system are more than 50% and less than 100% of 1 day man-labor cost.
Red: if the acummulated power losses of a PV system are equal or more than 1 day man-labor cost.

In this example 1 day man-labor cost = 1000 CHF and price per kWh = 0.6296 CHF.

NOTE: The icons for this optimizer have been created with this website.

1. Get initial location¶

The map will be started based on this initial location. In this case, we use Geneve as initial location city as follows:

Get location to initialize the map city = "Geneve" get location locator = geopy.geocoders.Nominatim(user_agent="MyCoder") location = locator.geocode(city) print('Location:' location) keep latitude and longitude only location = [location.latitude, location.longitude] print("[lat, long]:", location)

Location: Genève, 1204, Schweiz/Suisse/Svizzera/Svizra
[lat, long]: [46.2017559, 6.1466014]

2. Cumulative losses¶

The accumulated losses for SIG PV plants portfolio looks as follows:

3. Interactive map¶

As mentioned before:

Green: if the acumulated power losses of a PV system are not more than 50% of 1 day man-labor cost.
Orange: if the acumulated power losses of a PV system are more than 50% and less than 100% of 1 day man-labor cost.
Red: if the acummulated power losses of a PV system are equal or more than 1 day man-labor cost.

4. Clustering PV systems¶

The clustering of the PV systems is based on K-Means algorithm from sklearn library, with a twist. Basically what we do is the following:

K-Means clusters the PV systems based on the distance between them.
Once the clusters are calculated, we check for 2 different constrains:
1. Total 'Time to fix' of the entire cluster should not be more than 1 day.
2. Total 'PV energy losses' should be equivalent or higher than 1 day man-labor cost.

This cycle is repeated until the optimum number of clusters are defined. Same color circles indicate same cluster.